[Beowulf] Java vs C++ for interfacing to parallel library

Joe Landman landman at scalableinformatics.com
Sun Aug 20 13:24:29 PDT 2006



Geoff Jacobs wrote:

> Biased? Perhaps.
> 
> I think the compressed syntax as well as the broad semantics, which are
> a benefit to the implementer, really hamper ongoing maintenance of
> software written in Perl.

To avoid going into a language war, will basically leave this comment
with one note.  Most of the criticism of languages like Perl come from
groups that either don't or can't use them.  Javan's criticize
everything that is not Java, possibly not cognizant that the language of
their choosing is simply an exercise in corporate marketing.
OOo-ophiles criticize everything but the one true OO language ... which
ever that may be ... Pythonians criticize Perl as saying it is "too
hard", Rubians are usually escaped Pythonians and Javans saying how
wonderful it is compared to Perl (without bothering to note the
extremely strong similarities between the language syntax).  That is,
for any language, you will find detractors.  The detractors I pay
attention to are the ones who use it for a living or for their work.
The critics to ignore are the ones advocating another language.  That is
marketing, pure and simple.

I take a pragmatic view.  If a language is well suited for a task, I
think it makes sense to use it.  If the language would hinder the task,
why is it being considered over other superior platforms for expression?

I don't advocate against Fortran.  It is a great language, very useful
for its tasks.  It is hard to make it talk to other things, and it is 52
years old.  But when I teach an HPC programming class and I talk about
it, students audibly groan and tell me how their CS professors have told
them to ignore it.  Then again, most of the kids seem to know how to
program Java, C++, and VB.  A few even know Matlab.

The HPC User Forum and other folks note that lack of progress in HPC
software is impeding end users usage.  Anyone want to guess why?

To bring this back to the topic, I would suggest that the original
poster share more information on what they want to do.  Re-inventing
wheels isn't a great thing.  And the community here might be able to
help them find the previously invented wheels.

-- 
Joseph Landman, Ph.D
Founder and CEO
Scalable Informatics LLC,
email: landman at scalableinformatics.com
web  : http://www.scalableinformatics.com
phone: +1 734 786 8423
fax  : +1 734 786 8452 or +1 866 888 3112
cell : +1 734 612 4615



More information about the Beowulf mailing list